SIGHMY: Transforming Office Sighs into Positive Energy
Sighmy is an innovative, cactus-shaped device designed to transform workplace sighs—a common sign of stress and fatigue—into a playful and positive experience. By detecting sighs in real time through advanced audio processing and AI, Sighmy reacts with subtle movements that dissipate tension and uplift the office atmosphere.
2 weeks
Research-based Project
Product Designer
Machine Learning and Hardware Developer
Interviewer and Researcher
Data Analysis
Challenge
In modern workplaces, emotional well-being is often overlooked, yet it plays a crucial role in productivity and collaboration. Sighing—a subtle but frequent emotional signal—is often dismissed despite its ties to stress, fatigue, and dissatisfaction. As lots of research shows, a high frequency of sighs can dampen office morale and efficiency, yet few tools address this nuanced form of emotional expression.
What did I learn?
This project provided invaluable experience in designing solutions that cater to both business (toB) and consumer (toC) audiences. Through this, I honed my ability to conduct in-depth user research, synthesize diverse requirements, and translate them into a tangible user interface (TUI) and AI-powered system.
This project also reinforced my interest in designing emotionally intelligent technologies that support well-being and productivity, equipping me with a broader perspective on how to create systems that bridge personal and organizational goals effectively.
Inspiration
About Sigh
People produce about 12 "spontaneous sighs" per hour. These are long, deep breathing reflexes that usually include a second inhale and a subsequent exhale.
Sighing is a manifestation of emotion. By taking a deep breath and exhaling slowly, individuals can express negative emotions such as fatigue, frustration or anxiety. It serves as a way to vent emotions, help relieve inner pressure and regulate emotional states. Sighs not only reflect psychological and emotional states, but also involve physiological reactions, such as adjustments in breathing patterns, and serve as signals in social environments.
Background Research
Sighing Affects Work
Emotional Issues Become Prominent in Offices
Over half of workers feel worse in 2023 than in 2022, seeking understanding, not sympathy, from employers. Mood improved for those under 24, worsened for ages 25-29, and those over 30 face the most family and work pressure.
Insufficient Focus on Employees' Emotional Issues
Knowing Content — Primary Research
Employee Survey
Sample Size: 57 peoples; Age: 19 - 46+
The survey was designed based on existing research to understand the correlation between sighs, work atmosphere, and their intensity. A total of 57 questionnaires were collected. The survey results are divided into five parts, each highlighting key insights.
Survey Findings
Finding1: Reasons for Sighing During Work
This question is a multiple-choice question. Among them, 63.79% of people choose physical fatigue or discomfort, 48.28% are dissatisfied with the current situation, 51.72% lack resources or support, 51.72% have poor communication, and 56.9% have high work pressure.
Finding2: Emotions After Hearing Colleagues' Sighs
Only a very small percentage of people choose happy and angry, with the most common being calm, followed by worried, and then sad and frustrated. This indicates that others' sighs mainly bring a neutral or negative emotion.
Finding3: Frequency of Hearing Colleagues' Sighs and Office Atmosphere
From the chart, it can be observed that an office atmosphere with minimal sighs is better. As the frequency of sighs from people around increases, the office atmosphere tends to decline.
Finding4: Frequency of Sighing and Workload
From the chart, it can be seen that changes in workload before moderate levels do not cause too much variation in sighing frequency, while changes in workload after moderate levels are positively correlated with sighing frequency.
User Interviews
Emotional Response to Sighing: Participants reported feeling neutral or negative emotions, such as worry or frustration, when hearing colleagues sigh.
Impact on Atmosphere: Frequent sighing was observed to lower the overall office atmosphere and morale.
Perceived Causes of Sighing: Sighing was primarily triggered by heavy workload, poor communication, and lack of support.
Behavioral Observations: Sighing served as both a personal emotional outlet and a subtle signal influencing team dynamics.
Suggestions for Improvement: Participants suggested fostering a positive atmosphere and providing tools to manage stress without disrupting others.
Experimental design
Experiment Overview
Experiment 1: Task Load and Sigh Frequency
We measured the number of sighs per three-minute period for employees with varying job loads. Our results showed an upward trend in sigh frequency during multitasking.
Experiment 2: Vibration, Sound, Light, Images and their effects
Objective: To identify the most effective output modality (sound, touch, visual) for reducing stress and ensuring minimal disruption in the office and thus guide our next product design.
Experimental Design: During the completion of the task, participants were subjected to interference using vibration, sound, light, and image. After completing the task, participants evaluated the degree of interference they experienced.
Product Design
Feature Design
To Consumer (Employees)
Sigh frequency monitoring
Employee satisfaction comparison
Daily, weekly and monthly report summary
Work efficiency (time and task volume evaluation)
Work allocation arrangement suggestions
Employee growth trajectory (quarterly)
To Business (Employers)
Collect employee sigh emotions
Monitor employee task volume
Monitor working hours
Transform negative emotions into positive emotions
Data feedback
Collect employee satisfaction
Design Framework
Cactus
Cactus are common greenery in offices because of their ability to absorb radiation. So the image of it as as absorbing negative energy like sighs fits everyone's perception.
Inflatable Toy
Inflatable dolls will make people laugh because of their swinging funny image, thus improving the atmosphere of the office.
The Product — Sighmy
Sighmy (Sigh+Yummy) is a cactus-shaped smart device that absorbs negative emotions by responding to sighs in the office. With every sigh detected, it wiggles and disperses tension.
Sighmy increases the wind power as the number of sighs increases, ultimately making the plant wiggle happily. In this way, we hope the negative emotions that sighing brings to others are dissipated, and sighing is shown to be an effective way to relieve stress and encourage appropriate sighing.
Technology Implementation
Machine Learning Model
We used machine learning algorithms to train artificial intelligence (AI) models that can recognize sighs through the Edge Impulse platform.
Data Acquisition
Collected sighs and noise
Divided the training set and test set in the ratio of 8:2
Iterate the data acquisition process by expanding the origin data samples from S = 2 to S = 12.
Create Impulse (MFCC)
Use Mel Frequency Cepstral Coefficients (MFCC) to extract features from audio signals (see DSP results)
Classifier
MFCC as input features to train the neural network classifier
The accuracy of the training result is 90.8%
F1 score in Confusion Matrix for recognizing sighs is 0.92
Model Testing
The accuracy of the testing result is 89.30%
The final generalizability of the model meets our expectations
Hardware Design
Component Required
Data Transmission and Communication
The AI recognition model is deployed on Arduino nano 33 BLE board to detect the sigh, and then used Inter-integrated circuit (I2C) protocol to send data on the Arduino Uno board. When the UNO board received the signal of receiving a sigh
Circuit Diagram
Code in Arduino IDE
Prototype Testing Video
Final Product Showcase
Interface Design (To Employers)
In order to better analyze and guide the management activities of enterprise managers, a plugin interface based on collaborative work software Lark has been designed.
In addition to recording sigh data, this plugin can also use AI to infer employees' real-time anxiety and satisfaction levels, infer employee job saturation based on information obtained on Lark, and automatically generate reports and management suggestions.
